I’m curious about a couple Python stylistic decisions I’ve encountered.
The sets module has been deprecated since Python 2.6 in favor of native set() support. Is sets in use here because you need to support an older version of Python?
point_in_poly appears to implement a native algorithm instead of relying GEOS, OGR, or Shapely. Are those libraries unavailable where this code is being run?
python changewithin.py
Traceback (most recent call last):
File "changewithin.py", line 49, in
aoi_href = config.get('area', 'geojson')
File "/usr/lib/python2.7/ConfigParser.py", line 607, in get
raise NoSectionError(section)
ConfigParser.NoSectionError: No section: 'area'
We'll need to build in some safety guard for very voluminous changes. Should be incredibly cheap in the first iteration. Like just saying after X changes "There are (TOTAL - X) more changes not captured in this notification"
Right now changewithin flags any change to a building way. So if you add for instance a node on a straight way in a manner that does not change the actual shape of the way changewithin would still flag the change.
This leads to a lot of false positives for instance where user robgeb adds 3d details to buildings that actually don't have an impact on the building outline.
Can we deploy this change detection script on benzene? I know @tmcw has reached out to you already, but we'd love to set this up on benzene for testing.